I actually think this is pretty neat.

In order for this to pass tests, we'll need to:

• release python-graphblas, since this depends on unary.cbrt Done!
• update some NetworkX tests to use np.testing.assert_allclose when comparing floating point numbers (in progress)
  • while we're at it, it may be nice to add more tests to increase our coverage (in progress)
  • update CI to install NetworkX from github

With this, clustering.py is one step closer to being complete. It still needs: ALL DONE!

• square_clustering, no idea how to do this (but I haven't looked into it either) Done!
• generalized_degree, which looks very doable Done!